This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
[PATCH] Work around for GCC bug in frame.c
- From: Mark Kettenis <kettenis at chello dot nl>
- To: gdb-patches at sources dot redhat dot com
- Date: Sat, 8 May 2004 01:21:41 +0200 (CEST)
- Subject: [PATCH] Work around for GCC bug in frame.c
This patch works around a bug with GCC on some vax and m68k systems.
Andrew suggested introducing frame_eq() and frame_inner() and I
initially though that would be a good idea. However, there are almost
no places where these functions would be used, so I dropped the idea.
I've now committed my origional patch.
Mark
Index: ChangeLog
from Mark Kettenis <kettenis@gnu.org>
* frame.c (get_prev_frame_1): Work around GCC bug.
* frame.h: Add comment about GCC bug.
Index: frame.c
===================================================================
RCS file: /cvs/src/src/gdb/frame.c,v
retrieving revision 1.178
diff -u -p -r1.178 frame.c
--- frame.c 1 May 2004 22:41:34 -0000 1.178
+++ frame.c 7 May 2004 23:13:48 -0000
@@ -1609,6 +1609,7 @@ static struct frame_info *
get_prev_frame_1 (struct frame_info *this_frame)
{
struct frame_info *prev_frame;
+ struct frame_id this_id;
gdb_assert (this_frame != NULL);
@@ -1646,7 +1647,8 @@ get_prev_frame_1 (struct frame_info *thi
/* Check that this frame's ID was valid. If it wasn't, don't try to
unwind to the prev frame. Be careful to not apply this test to
the sentinel frame. */
- if (this_frame->level >= 0 && !frame_id_p (get_frame_id (this_frame)))
+ this_id = get_frame_id (this_frame);
+ if (this_frame->level >= 0 && !frame_id_p (this_id))
{
if (frame_debug)
{
@@ -1663,16 +1665,14 @@ get_prev_frame_1 (struct frame_info *thi
go backwards) and sentinel frames (the test is meaningless). */
if (this_frame->next->level >= 0
&& this_frame->next->type != SIGTRAMP_FRAME
- && frame_id_inner (get_frame_id (this_frame),
- get_frame_id (this_frame->next)))
+ && frame_id_inner (this_id, get_frame_id (this_frame->next)))
error ("Previous frame inner to this frame (corrupt stack?)");
/* Check that this and the next frame are not identical. If they
are, there is most likely a stack cycle. As with the inner-than
test above, avoid comparing the inner-most and sentinel frames. */
if (this_frame->level > 0
- && frame_id_eq (get_frame_id (this_frame),
- get_frame_id (this_frame->next)))
+ && frame_id_eq (this_id, get_frame_id (this_frame->next)))
error ("Previous frame identical to this frame (corrupt stack?)");
/* Allocate the new frame but do not wire it in to the frame chain.
Index: frame.h
===================================================================
RCS file: /cvs/src/src/gdb/frame.h,v
retrieving revision 1.132
diff -u -p -r1.132 frame.h
--- frame.h 1 May 2004 22:41:34 -0000 1.132
+++ frame.h 7 May 2004 23:13:48 -0000
@@ -310,7 +310,22 @@ extern CORE_ADDR get_frame_base (struct
/* Return the per-frame unique identifer. Can be used to relocate a
frame after a frame cache flush (and other similar operations). If
- FI is NULL, return the null_frame_id. */
+ FI is NULL, return the null_frame_id.
+
+ NOTE: kettenis/20040508: These functions return a structure. On
+ platforms where structures are returned in static storage (vax,
+ m68k), this may trigger compiler bugs in code like:
+
+ if (frame_id_eq (get_frame_id (l), get_frame_id (r)))
+
+ where the return value from the first get_frame_id (l) gets
+ overwritten by the second get_frame_id (r). Please avoid writing
+ code like this. Use code like:
+
+ struct frame_id id = get_frame_id (l);
+ if (frame_id_eq (id, get_frame_id (r)))
+
+ instead, since that avoids the bug. */
extern struct frame_id get_frame_id (struct frame_info *fi);
extern struct frame_id frame_unwind_id (struct frame_info *next_frame);